GaAs semiconductor technology provides transistor devices of the npn bipolar transistor and n-FET types. Circuit design and system functionality require p-type devices, especially pnp bipolar transistors or p-channel field-effect transistors, which are not provided in the GaAs technology up to now. With only n-type devices available in GaAs technology, other circuit concepts are applied at the expense of increased current consumption, larger occupied area and other shortcomings. Semiconductor circuits in silicon comprise both n-type and p-type devices, but this is not the preferred technology for power amplifiers, especially in industrial applications. Although the lacking p-type transistor components can be provided on external circuit components, a full integration of both types of transistors is desired also in GaAs technology.

Merged or stacked FET-HBT integration schemes, often called BiFET or BiHEMT and containing both HBT and FET or P-HEMT devices on a single GaAs substrate, are reported in the following papers from the CS MANTECH Conference 2007: William Peatman, Mohsen Shokrani, Boris Gedzberg, Wojciech Krystek, and Michael Trippe: "InGaP-Plus™: Advanced GaAs BiFET Technology and Applications";
T. Henderson, J. Middleton, J. Mahoney, S. Varma, T. Rivers, C. Jordan, and B. Avrit: "High-Performance BiHEMT HBT / E-D pHEMT Integration";
Todd D. Basso and Richard B. Brown: "A Complementary GaAs Microprocessor for Space Applications";
Ravi Ramanathan, Mike Sun, Peter J. Zampardi, Andre G. Metzger, Vincent Ho, Cejun Wei, Peter Tran, Hongxiao Shao, Nick Cheng, Cristian Cismaru, Jiang Li, Shiaw Chang, Phil Thompson, Mark Kuhlman, Kenneth Weller: "Commercial Viability of a Merged HBT-FET (BiFET) Technology for GaAs Power Amplifiers"; C. K. Lin, T. C. Tsai, S. L. Yu, C. C. Chang, Y. T. Cho, J. C. Yuan, C. P. Ho, T. Y. Chou, J. H. Huang, M. C. Tu, and Y. C. Wang: "Monolithic Integration of E/D-mode pHEMT and InGaP HBT Technology on 150-mm GaAs Wafers";

US 6,326,317 B1 describes an enhancement-mode p-MOSFET on a GaAs substrate comprising a mesa formed by an undoped epitaxial buffer layer, a moderately doped n-type layer, and a heavily doped p⁺-type layer. A gate electrode is arranged on an oxide layer within a recess of the p⁺-type layer.

It is an object of this invention to provide a p-channel field-effect transistor in GaAs technology. A further object is the disclosure of an efficient production method of a p-channel field-effect transistor in GaAs technology.

These and other objects are achieved by the p-type field-effect transistor according to claim 1 and the method of production according to claim 6, respectively.

The field-effect transistor is based on a layer structure that can be used to manufacture npn bipolar transistors. A p-doped layer, which can also serve as base layer of a bipolar transistor, is divided into two separate portions by a recess, in which a gate electrode is located. The gate electrode is on a gate dielectric layer and controls the space charge in the channel in an n-doped layer, which can also serve as collector layer of a bipolar transistor. Thus, the p-channel field-effect transistor can be arranged in a layer sequence that can be used in a vertical npn bipolar transistor. Furthermore, this layer sequence is laterally confined to form a mesa on a substrate surface. An n⁺-doped layer, which can also serve as subcollector layer of a bipolar transistor, is positioned under the mesa and is provided with an external electric connection functioning as a bulk contact of the field-effect transistor.

Examples of the field-effect transistor and the production method are described in more detail in conjunction with the appended figures.

FIG. 1

shows a cross-section of an embodiment of the p- FET.

FIG. 2

shows a schematic plan view of the electric contacts and connections of the p-FET.

FIG. 1 shows a cross-section of an embodiment of the p-channel field-effect transistor. An n⁺-layer 2 is located on an upper surface of a substrate 1, which can be especially GaAs. On the n⁺-layer 2, a layer sequence is arranged forming a mesa, which comprises an n-layer 3, a p-layer 4 and a further n-layer 5. This layer sequence can be provided for npn bipolar transistors, which can be manufactured on the same substrate. Thus, the field-effect transistor can use the layer sequence of a standard GaAs HBT manufacturing technology. The upper n-layer 5, which is shown in the cross-section of FIG. 1, is not necessary for the field-effect transistor. It could therefore be left out or removed, but it can be left on the p-layer 4 as well if the layer sequence as shown in FIG. 1 is provided in the production method.

The p-layer 4 is divided by a trench or recess into two separate portions as shown in FIG. 1. The bottom and sidewalls of the recess are covered with a relatively thin gate dielectric 6. Optionally, source/drain contact layers 7, preferably a metal, may be arranged on the surface of the portions of the p-layer 4. The regions that are occupied by space charge regions 8 during the operation of the transistor are indicated by the hatched areas in FIG. 1. A channel region 9 is provided under the gate dielectric 6 within the lower n-layer 3 between the space charge regions 8. Since the channel region 9 is located in n-doped semiconductor material, it is provided for a p-channel field-effect transistor. A gate electrode 10, which can be a metal, is arranged in the recess on the gate dielectric 6. Source/drain contacts 11 are provided for source and drain and, in this embodiment, arranged on the source/drain contact layers 7. The source/drain contacts 11 are preferably a metal.

The n⁺-layer 2 can be provided as a subcollector layer of a vertical npn transistor. In the field-effect transistor, the n⁺-layer 2 can be used to provide a bulk contact. To this purpose, a bulk contact layer 12, preferably a metal, is arranged on the upper surface of the n⁺-layer 2 in an area that is not covered by the n-layer 3. A bulk contact 13, which is preferably metal, is applied to the bulk contact layer 12. An isolating layer 14 on the level of the n⁺-layer 2 can be provided as a lateral insulation of the n⁺-layer 2, and a further dielectric layer 15 can be provided as a passivation and electric insulation between the various electric contacts on the upper side of the device. The arrangement and shape of the layers, especially the gate dielectric 6, the dielectric layer 15, and the contact layers, may deviate from the example of FIG. 1 and can be varied in order to optimize the performance of the field-effect transistor.

The field-effect transistor is especially advantageous in that the source/drain regions are formed by portions of the p-layer 4, which can also serve as the base layer of a vertical npn bipolar transistor, and the channel region is located in an n-layer 3, which can also serve as a collector layer of an npn bipolar transistor. The source/drain regions are completely separated from one another by the recess comprising the gate electrode and the gate dielectric. Thus there is no direct connection via p-doped semiconductor material between the two portions of the p-layer 4. In this manner, a sequence of p-doped, n-doped and p-doped semiconductor layers can be used to form a field-effect transistor of essentially lateral structure.

FIG. 2 shows the layout of the electrically conducting tracks in a schematic plan view. The gate electrode 10 is shown to be arranged between the source/drain contacts 11, while the bulk contact 13 is arranged laterally to the other contacts. In this example every contact 10, 11, 13 is electrically conductively connected with a contact pad 16, which can be used to apply an external electric connection. But the design of the conducting tracks and contact layers can be varied according to the requirements of each individual embodiment. Thus there is a large variety of possibilities to integrate the field-effect transistor together with other devices, especially vertical bipolar transistors.

The field-effect transistor can be manufactured within a standard GaAs technology together with vertical bipolar transistors. The method of production of the field-effect transistor comprises the steps of applying an n-layer and a p-layer above a substrate of semiconductor material, especially of GaAs, the layers being preferably structured in the shape of a mesa. A recess or trench is etched into the p-layer in order to form two separate portions, which are not connected by p-doped semiconductor material. A gate dielectric is arranged on the bottom and sidewalls of the recess. The gate dielectric can be silicon nitride, which can be used within a standard GaAs HBT technology. A gate electrode, preferably formed from metal, is arranged in the recess on the gate dielectric. Source/drain contacts are arranged in electric connection with the portions of the p-layer forming source and drain. An n⁺-layer is arranged between the substrate and the mesa comprising the n-layer and the p-layer. The n⁺- layer is provided with a bulk contact and with a bulk contact layer between the bulk contact and the n⁺-layer.

The method can be used to produce also vertical bipolar transistors. To this end, a further n-layer is arranged on the p-layer. The n⁺-layer is used as a subcollector layer, and the lower n-layer, the p-layer and the upper n-layer are used in this sequence as collector layer, base layer and emitter layer of a vertical npn bipolar transistor. The layer sequence is structured, especially by forming mesas for each individual device, and the field-effect transistor is produced by forming the recess that is intended for the gate electrode.

Thus only a slight modification of the standard production process of vertical bipolar transistors in GaAs technology enables the production of p-channel field-effect transistors in this semiconductor material. The production method is therefore compatible with standard GaAs HBT technology, does not require essential variations of the process steps and masks and is therefore highly cost-effective. An optimization of the electric properties of the field-effect transistor is easily achieved by an adaptation of the dimensions of the gate electrode, the thickness of the gate dielectric and the arrangement of the contact layers.

1

substrate

2

n⁺-layer

3

n-layer

4

p-layer

5

further n-layer

6

gate dielectric

7

source/drain contact layer

8

space charge region

9

channel region

10

gate electrode

11

source/drain contact

12

bulk contact layer

13

bulk contact

14

isolating layer

15

dielectric layer

16

contact pad